INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The nucleus accumbens (NAcc) is a subregion of the ventral striatum that plays a critical role in reinforcement and reward (Kelley et al. 1997; Salamone 1996). Activity in the NAcc has been suggested to contribute to motivational aspects of drug-seeking behavior, playing a role in both the initial rewarding aspects of drug use and the formation of psychological dependence and withdrawal (Koob et al. 1992, 1998; Nestler 1996; Stinus et al. 1990). The NAcc can be divided into core and shell subregions based on anatomical and biochemical differences (Jongen-Relo et al. 1993, 1994; Meredith et al. 1993; Voorn et al. 1994; Zaborszky et al. 1985), and it has been suggested that the core is part of the striatal complex; however, the shell can be considered a component of the extended amygdala (Alheid and Heimer 1988; Heimer et al. 1997). There is mounting evidence that the core and the shell have different functional roles. The administration of drugs of abuse has been shown to cause an increase in dopamine release, selectively, in the shell (Di et al. 1993; Pontieri et al. 1995), and animals will self-administer drugs of abuse into the shell but not the core (Carlezon and Wise 1996a,b; Carlezon et al. 1995). It has thus been suggested that the shell may be involved in stimulus-reward associations and that this system is altered by drugs of abuse (Koob et al. 1998; Wise 1996; Zahm 1999). On the other hand, the core has been shown to play a critical role in conditioning models of drug-seeking behavior, such as the Pavlovian approach (Di Ciano et al. 2001; Everitt et al. 2001), and in cue-associated drug-seeking behavior, there is a differential effect of N-methyl-D-aspartate (NMDA) and -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)-kainate antagonists between the core and the shell (Di Ciano and Everitt 2001). The MSNs in different compartments of the NAcc differ in their afferent and efferent connections and thus in their physiological and behavioral functions.

The primary output neurons of the NAcc are GABAergic MSNs, which project to various regions within the mesencephalon, basal ganglia, and extended amygdala (Groenewegen and Russchen 1984; Heimer et al. 1991; Kalivas et al. 1993; Zahm and Brog 1992). These cells rest at a negative membrane potential and rely on excitatory glutamatergic input to drive them to the threshold for firing action potentials (Uchimura et al. 1989; Wilson and Kawaguchi 1996; Wilson et al. 1983). The specific timing and frequency of firing is further modulated by inputs from GABAergic and cholinergic interneurons within the NAcc (Galarraga et al. 1999; Kawaguchi et al. 1995; Koos and Tepper 1999). Hence excitatory and inhibitory synaptic inputs ultimately regulate the processing and output of this brain region.

The synaptic inputs that regulate MSN activity are themselves modulated by a number of presynaptic signaling pathways. Three such pathways that have received considerable attention for their potential role in reinforcement and reward are opioid, adenosine receptors, and cAMP. There is compelling evidence that these signaling pathways interact with one another. Opioids acutely inhibit adenylyl cyclase activity (Heijna et al. 1992; Izenwasser et al. 1993), and chronic opioid treatment can upregulate cAMP formation (Avidor-Reiss et al. 1995; Sharma et al. 1975; Terwilliger et al. 1991). Furthermore, cAMP can be converted extracellularly into adenosine, and the inhibitory effects of adenosine on synaptic activity may act as a negative feedback modulator for the excitatory effects of cAMP-dependent protein kinase (PKA) (Bonci and Williams 1996; Dunwiddie and Hoffer 1980; Lu and Gean 1999; Rosenberg et al. 1994). The strength of the synaptic inputs to NAcc MSNs is thus regulated by a complex interaction between these neurochemical signals.

Although the actions of opioid receptors, adenylyl cyclase-cAMP, and adenosine have been characterized in several systems, the ability of these agents to modulate synaptic activity has not been rigorously compared across different synapses within the NAcc. The present study characterizes the function of opioid receptors, adenylyl cyclase-cAMP, and adenosine at a single isolated response: AMPA receptor-mediated excitatory postsynaptic currents (EPSCs) in the shell subregion of the NAcc. This study provides a comparison of the ability of each of these signaling pathways to modulate synaptic activity at four different synapses in the NAcc: EPSCs and inhibitory postsynaptic currents (IPSCs) recorded from MSNs of the core and shell. The major finding of this study is that under nearly identical recording conditions, the ability of these signaling pathways to modulate synaptic activity varies across the different synapses of the NAcc.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Preparation of NAcc slices

Young male Wistar rats (140-160 g) were anesthetized with a ketamine-xylazine-acepromazine mixture (50:2:1 mg/kg) and killed, and their brains were rapidly removed and cut into 250-µm horizontal slices with a vibratome at 4°C. Tissue surrounding the NAcc was removed, and the slices were stored in physiological saline at room temperature. Only slices that contained a continuous layer of white matter from the anterior commissure along the lateral side of the NAcc were used. In this way slices were selected from midway between the dorsal and ventral ends of the NAcc and consistently contained equal portions of core and shell (Paxinos and Watson 1998; Fig. 1). The saline used in all experiments contained the following (in mM): 126 NaCl, 2.5 KCl, 1.2 MgCl₂, 2.4 CaCl₂, 1.2 NaH₂PO₄, 11 glucose, and 21.4 NaHCO₃, oxygenated with 95% O₂-5% CO₂. The incubation solution also contained the NMDA antagonist (RS)-3-(2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid (CPP; 10 µM) to prolong the health of the slices. After 1-4 h, the slices were transferred to a recording chamber and superfused with physiological saline at 2 ml/min.

View larger version (77K): [in this window] [in a new window]   Fig. 1. Image of the brain regions recorded. A: horizontal section indicating the region of the nucleus accumbens (NAcc) used in this study. Section is taken from the plane at Bregma 7.10 mm (Paxinos and Watson 1998). Slices were within 250 µm dorsoventral of this plane. Diagram adapted (Paxinos and Watson 1998). B: low-power image of a NAcc slice indicating regions in which neurons from the shell and core were recorded. Stimulation was always rostral to the site of recording.

Electrophysiological recording

Whole cell recordings were made from MSNs with an Axopatch 1D patch-clamp amplifier. Recording electrodes were pulled from borosilicate glass (OD 1.5 mm, ID 1.2 mm, with filament; World Precision Instruments) on a Narishige micropipette puller and had tip resistances of 2-4 M when filled with a solution containing the following (in mM): 125 Cs-gluconate, 11 KCl, 10 HEPES, 0.1 CaCl₂, 1 K-EGTA, 2 Mg-ATP, 0.3 Tris-guanosine-5-triphosphate (GTP), pH adjusted to 7.3 with KOH, osmolarity adjusted to 288 mOsm. Cells were visualized with a 40× water immersion lens with Normarski optics and infrared illumination. MSNs were selected within 200 µm of either the medial boundary (shell) or the lateral boundary (core) of the NAcc (Paxinos and Watson 1998; Fig. 1). MSNs were identified based on their small size (10-15 µm diameter), negative resting membrane potential (approximately 75 to 80 mV), lack of spontaneous action potentials, and the presence of a fast inwardly rectifying potassium current in the absence of a slow hyperpolarization-activated current (I_h) at negative membrane potentials (Uchimura et al. 1989). Synaptic currents were evoked every 20 s with a tungsten bipolar stimulating electrode (Frederick Haer and Company) placed on the surface of the slice rostral to the recording electrode. EPSCs were recorded by voltage-clamping the MSNs at 75 mV in the presence of CPP (10 µM) and picrotoxin (100 µM). The resulting inward EPSCs were completely blocked by application of the AMPA receptor antagonist 6-nitro-7-sulfamoylbenzo[f]quinoxaline-2,3-dione (NBQX; 5 µM). IPSCs were recorded at 5 mV in the presence of CPP (10 µM) and NBQX (5 µM). These outward IPSCs were completely blocked by application of the GABA_A receptor antagonist picrotoxin (100 µM). The stimulation intensity for evoked currents (both EPSCs and IPSCs) was generally adjusted to produce a response in the 400-800 nA range. This range was always submaximal and allowed the currents to move in either direction without reaching a floor or ceiling.

Evoked synaptic currents were acquired and analyzed with Acquis1 software version 4.0 (Bio-logic SA, Grenoble, France). The amplitude of each evoked event was determined, and measurements from each cell were made by averaging 10 events obtained during the last 3 min of the baseline or each drug administration. Spontaneous miniature synaptic currents were recorded as above except no stimulating electrode was used and tetrodotoxin (500 nM) was included in the bath. Miniature synaptic currents were acquired with pClamp 6.0 (Axon Instruments) by recording a 2-s sweep every 10 s and were analyzed with Axograph 4.0 as follows. For each cell, all of the spontaneous synaptic events during a sample period of the baseline were selected by eye and the events averaged. The amplitude, rise, and decay of this average event were used to construct a variable amplitude template in Axograph. The detection threshold was then adjusted to allow Axograph to detect the same events as determined by eye for the sample baseline period, with minimal positive or negative errors. The template and threshold criteria were then used to detect events for the entire experiment. This procedure was repeated for each cell. The frequency was determined within each 2-s sweep, all events within each sweep were averaged, and the peak amplitude, rise, and decay time courses were measured. In this way, average measurements of frequency, amplitude, rise, and decay were determined every 10 s. The final measurements reported are the average of eighteen 10-s bins recorded during the last 3 min of the baseline or each drug administration.

Access resistance was determined with a bridge circuit in current-clamp mode at the beginning and end of each experiment and was below 20 M at all times. Series resistance was compensated by 80%, except for the experiments in which spontaneous miniature currents were recorded, in which no compensation was used to reduce electrical noise. The voltages reported have been corrected for a 15 mV liquid-liquid junction potential, as determined with JPCalc software (Barry 1994). All drugs were applied by dissolving them directly into the superfusion solution.

Statistical analysis

Statistical comparisons were made with Prism version 3 software (GraphPad). The paired two-tailed student's t-test was used to determine whether treatments produced a significant effect relative to the baseline recording. The Mann-Whitney test was used to compare normalized responses between cells. The criterion for statistical significance was P < 0.05.

Chemicals

[Met]⁵enkephalin acetate (ME), [D-Ala2-N-Me-Phe4-Gly5]-enkephalin (DAMGO), adenosine, and N6-cyclopentyladenosine (CPA) were obtained from Sigma Chemical (St. Louis, MO). Picrotoxin, forskolin, and 8-cyclopentyl-1,3-dipropylxanthine (DPCPX) were from Research Biochemicals International. CPP and NBQX were obtained from Tocris Cookson. D-Phe-Cys-Tyr-D-Trp-Arg-Thr-Pen-Thr-NH2 (CTAP) was from Phoenix Pharmaceutical.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Opioid inhibition of EPSCs in shell MSNs

The effects of opioids were initially examined on pharmacologically isolated AMPA receptor-mediated EPSCs in MSNs of the NAcc shell (Fig. 2). The opioid agonist ME (10 µM) inhibited excitatory synaptic transmission in all cells tested (mean 29 ± 3% inhibition; n = 11; P < 0.0001 by t-test; Fig. 2B). Because ME is known to act on both µ- and -opioid receptors, synaptic inhibition by the µ-selective peptide agonist DAMGO was examined. DAMGO caused a dose-dependent inhibition of EPSCs (maximum inhibition 30%; Fig. 2C). This inhibition was completely reversed by the µ-selective antagonist CTAP (1 µM; data not shown). Hence DAMGO produced an amount of inhibition similar to that of ME. In the presence of CTAP (1 µM), the -selective peptide agonist [D-Pen², D-Pen²]enkephalin (DPDPE) produced less consistent results, inhibiting EPSCs in two of five cells tested (data not shown). DPDPE (1 µM) thus produced an average of 16 ± 10% inhibition of shell EPSCs, which was not statistically significant (n = 5; P = 0.17 by t-test). It appears that µ-receptors are the primary mediators of opiate-induced inhibition of EPSCs, although -receptors may make a contribution in a subset of cells.

View larger version (15K): [in this window] [in a new window]   Fig. 2. Opioid inhibition of excitatory postsynaptic currents (EPSCs) in NAcc shell medium spiny neurons (MSNs). A: time course of effects of ME on a representative shell MSN. Horizontal bar, time of [Met]5enkephalin (ME) application. Each point on the time course is the average of 3 EPSCs. Each synaptic current trace is the average of 10 consecutive EPSCs taken at the time indicated by the arrow. B: average time course of effects of ME across all experiments (n = 11). C: dose-response curve for the Ala2-N-Me-Phe4-Gly5-enkephalin (DAMGO)-mediated inhibition of EPSCs recorded from shell MSNs. Each point is the average of 4-9 separate experiments. Curve represents the best fit to the data with the equation y = min + (max  min)/(1 + 10((LogEC50  x) × HillSlope)), with the minimum fixed at 0. Calculated maximum was 30.1%, calculated EC50 was 0.24 µM, and calculated Hill Slope was 1.55. Error bars, SE.

Forskolin potentiation of excitatory synaptic currents in shell MSNs

To determine how changes in cAMP levels modulate the activity of synapses in the NAcc shell, the effects of forskolin, an activator of adenylyl cyclase, were examined on shell EPSCs (Fig. 3). Forskolin (10 µM) augmented EPSCs by 41 ± 10% (n = 11; P < 0.01 by t-test; Fig. 2B). To assess the extent to which pre- and postsynaptic mechanisms account for the facilitation of evoked synaptic transmission, the paired-pulse ratio (i.e., the relative amplitude of two synaptic responses) was determined for shell EPSCs at a 50-ms interval. An increase in the paired-pulse ratio is usually associated with presynaptic inhibition, and a decrease in the ratio is associated with presynaptic potentiation. No change in the paired-pulse ratio suggests the effect is postsynaptic (Creager et al. 1980; Harris and Cotman 1983). The paired-pulse ratio decreased in every cell tested after application of forskolin (10 µM), from an average of 1.15 ± 0.10 to an average of 0.91 ± 0.07 (n = 6 cells), a significant decrease (P < 0.01 by paired t-test). This supports the idea that the increase induced by forskolin was mediated by a presynaptic mechanism. To further assess the relative contribution of pre- and postsynaptic mechanisms on the effects of forskolin, spontaneous miniature EPSCs (mEPSCs) were recorded from shell MSNs in the presence of tetrodotoxin (500 nM; Fig. 4). The AMPA receptor antagonist NBQX (5 µM) reduced the frequency of mEPSCs to zero, indicating that all of the responses detected were AMPA receptor-mediated synaptic events. The baseline frequency of mEPSCs was 9.5 ± 3.0 Hz, and the baseline amplitude was 21.8 ± 2.2 pA.. Forskolin increased the frequency of mEPSCs by 232 ± 67% (P = 0.002 by paired t-test) and increased the amplitude by 31 ± 5% (n = 4; P = 0.02 by paired t-test). The increase in both the frequency and amplitude of the response suggests that forskolin may act through both pre- and postsynaptic mechanisms. However, the much larger effect of forskolin on the frequency suggests that presynaptic mechanisms are likely to account for most of the increase seen in the evoked response.

View larger version (21K): [in this window] [in a new window]   Fig. 3. Forskolin potentiation of EPSCs in NAcc shell MSNs. A: time course of forskolin effects on a representative shell MSN. Horizontal bar, time of forskolin application. Each point is the average of 3 EPSCs. Each trace is the average of 10 consecutive EPSCs taken at the time indicated by the arrow. B: average time course of forskolin effects across all experiments (n = 11). Error bars, SE.

View larger version (29K): [in this window] [in a new window]   Fig. 4. Forskolin increased both frequency and amplitude of spontaneous miniature EPSCs (mEPSCs) in shell MSNs. Top: traces from representative cell showing spontaneous mEPSCs before and after application of forskolin. Graphs are a summary of 4 experiments showing both the frequency and amplitude of mEPSCs were increased by forskolin. Error bars, SE.

Adenosine inhibition of excitatory synaptic currents in shell MSNs

The sensitivity of EPSCs to inhibition by adenosine was examined in shell MSNs (Fig. 5A). Both adenosine and the A₁ adenosine receptor agonist CPA inhibited shell EPSCs in a dose-dependent manner, and this inhibition could be completely reversed by the A₁ receptor antagonist DPCPX (200 nM), demonstrating that shell EPSCs are highly sensitive to the inhibitory effects of A₁ adenosine receptors. Many regions of the brain are subject to a tonic low-level inhibition by endogenous adenosine (Ballarin et al. 1991; Dunwiddie and Diao 1994; Dunwiddie and Hoffer 1980). The magnitude of this tonic inhibition can be determined by measuring the potentiation of synaptic currents during application of an adenosine receptor antagonist (Brundege and Dunwiddie 1996, 1998). To determine the tonic inhibition mediated by endogenous adenosine in the NAcc, we applied the adenosine A₁ receptor antagonist DPCPX (100 nM; Fig. 5B). DPCPX failed to produce a significant increase in shell EPSCs (n = 6; 4 ± 4% increase; P = 0.33 by two-tailed t-test), suggesting that there is very little endogenous adenosine present in the vicinity of glutamatergic inputs to shell MSNs.

View larger version (23K): [in this window] [in a new window]   Fig. 5. Adenosine A1 receptors inhibit shell MSN EPSCs, but there is little tonic inhibition due to endogenous adenosine. A: inhibition of shell EPSCs by adenosine and the A1 receptor selective agonist N6-cyclopentyladenosine (CPA). Adenosine effects quickly wash out after superfusion is stopped. Effects of CPA were completely reversed by the A1 adenosine receptor antagonist 8-cyclopentyl-1,3-dipropylxanthine (DPCPX). Horizontal bars, addition of drugs at concentrations given. Each point is the average of three responses, and the synaptic currents shown are the average of 10 responses taken at the time indicated by the arrow. B: application of DPCPX caused a small nonsignificant increase in shell EPSCs, indicating little or no tonic inhibition of these synapses by endogenous adenosine. Error bars, SE.

Comparison of the effects of opioids between different synapses in the NAcc

The output of MSNs is determined through the net effect of both excitatory and inhibitory synaptic inputs, suggesting that a true evaluation of opioid activity requires an evaluation of opioid effects on both of these synapses. Furthermore, there may be functional differences between MSNs in different subregions of the NAcc, and these may receive synaptic inputs from different areas of the brain. Thus to gain a more complete picture of the effects of opioids and related signaling pathways on MSN activity, the effects of several compounds were compared between excitatory and inhibitory synapses in both shell and core MSNs.

To compare the effects of opioids between different synapses, the effects of ME (10 µM) were examined on both EPSCs and GABAergic inhibitory postsynaptic currents (IPSCs) in MSNs of both the shell and core subregions of the NAcc. Figure 6 shows examples of the inhibition caused by ME at the four synapses tested. Figure 7 compares the mean inhibition from each of these synapses. The data for the shell EPSC responses is the same data used in Fig. 2B and is shown again for comparison. ME (10 µM) significantly inhibited all four of the synapses tested (shell EPSCs 29 ± 3% inhibition, n = 11; core EPSCs 39 ± 4% inhibition, n = 5; shell IPSCs 68 ± 3% inhibition, n = 6; core IPSCs: 47 ± 9% inhibition, n = 6). On average, ME inhibited shell IPSCs to a significantly greater extent than each of the other synapses (vs. core IPSC P = 0.0411, vs. shell EPSC P = 0.0011; vs. core EPSC P = 0.0043 Mann-Whitney test), suggesting that inhibitory synapses in the shell may be particularly sensitive to the effects of opioids. There were no significant differences in the effects of ME between the other synapses.

View larger version (30K): [in this window] [in a new window]   Fig. 6. Examples of the inhibition by ME at different synapses in the NAcc. Summarized data are presented in Fig. 7. A-D: time courses of effects of ME (10 µM) on synaptic currents in a MSN. Horizontal bar, time of ME application. Each point on the time course is the average of 3 synaptic currents. Each example current trace shown above the time courses is the average of 10 consecutive currents taken at the time indicated by the arrow. IPSC, inhibitory postsynaptic current.

View larger version (59K): [in this window] [in a new window]   Fig. 7. ME differentially inhibits synapses in the NAcc. %Inhibition induced by ME (10 µM) at 4 different synapses in the NAcc. ME caused significant inhibition in all cells tested. Inhibition of shell IPSCs was significantly greater than at the other synapses. **Significant inhibition of synaptic response relative to baseline as determined by 2-tailed t-test.

Comparison of the effects of forskolin between different synapses in the NAcc

The effects of the adenylyl cyclase activator forskolin (10 µM) were also determined at inhibitory and excitatory synapses of both shell and core MSNs. Figure 8A shows the effects of forskolin on IPSCs in a typical shell MSN. A comparison of all four synapses is shown in Fig. 8B. Forskolin significantly potentiated synaptic currents at all four synapses tested (shell EPSCs 41 ± 10% increase, n = 11; core EPSCs 68 ± 10% increase, n = 4; shell IPSCs 78 ± 12% increase, n = 7; core IPSCs 56 ± 8% increase, n = 9).

View larger version (26K): [in this window] [in a new window]   Fig. 8. Forskolin potentiates synaptic currents at 4 different synapses in the NAcc. A: time course of the effects of forskolin on IPSCs from a representative shell MSN. Horizontal bar, time of forskolin application. Each point is the average of 3 IPSCs. Each synaptic current trace shown is the average of 10 consecutive IPSCs taken at the time indicated by the arrow. B: %inhibition induced by forskolin (10 µM) at 4 different synapses in the NAcc. **Significant potentiation of the synaptic response relative to baseline as determined by 2-tailed t-test.

The effects of forskolin at EPSCs in shell MSNs were shown to be primarily due to presynaptic activity, with some contribution from a postsynaptic effect (Fig. 4). To determine the presynaptic and postsynaptic effects of forskolin on inhibitory synaptic transmission, spontaneous miniature IPSCs (mIPSCs) were recorded from NAcc shell MSNs in the presence of tetrodotoxin (500 nM; Fig. 9). The baseline frequency of mIPSCs was 3.0 ± 0.4 Hz, and the baseline amplitude was 21.3 ± 0.8 pA. Picrotoxin (100 µM) reduced the frequency of these responses to zero, demonstrating that the responses detected were GABA_A receptor-mediated synaptic currents. Forskolin (10 µM) increased the frequency of mIPSCs by 112 ± 19% (n = 5; P = 0.0003 by paired t-test) but had no effect on the amplitude (amplitude after forskolin 96 ± 5% of control; P = 0.56 by paired t-test). These data suggest that forskolin increased the probability of neurotransmitter release at inhibitory synapses in the shell of the NAcc through a purely presynaptic mechanism and had no effect on postsynaptic GABA_A-mediated currents.

View larger version (27K): [in this window] [in a new window]   Fig. 9. Forskolin increased the frequency but not the amplitude of spontaneous miniature IPSCs (mIPSCs). Top: traces from a representative cell showing spontaneous mIPSCs before and after application of forskolin. Graphs are a summary of 5 experiments. Error bars, SE.

Comparison of endogenous adenosine levels between different synapses in the NAcc

The level of tonic inhibition by endogenous adenosine was determined at each of the four synapses by blocking A₁ adenosine receptors with the antagonist DPCPX (100 nM). As shown in Fig. 10, DPCPX had no significant effect on EPSCs in shell MSNs. However, DPCPX significantly increased synaptic currents at core EPSCs and at shell and core IPSCs (shell EPSCs 4 ± 4% increase, n = 6; core EPSCs 18 ± 5% increase, n = 5; shell IPSCs 15 ± 5% increase, n = 9; core IPSCs 34 ± 13% increase, n = 11). These results suggest that there is a significant amount of tonic adenosine at most of the synapses within the NAcc but that levels are particularly low in the region of excitatory synapses to MSNs of the shell.

View larger version (41K): [in this window] [in a new window]   Fig. 10. Differential levels of tonic inhibition by endogenous adenosine at NAcc synapses. Tonic inhibition of synapses by endogenous adenosine as revealed by application of the adenosine A1 receptor antagonist DPCPX (100 nM). DPCPX caused a significant increase in synaptic currents at 3 synapses but caused no significant increase in shell EPSCs. *Significant potentiation of synaptic response relative to baseline as determined by 2-tailed t-test.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The overall goal of this study was to compare the actions of drugs acting at three signaling pathways across four different synapses within the NAcc. ME and adenosine inhibited, whereas forskolin potentiated, synaptic transmission at all four synapses. However, there were differences in the pattern of modulation between these synapses. ME varied in the magnitude of inhibition. Forskolin varied in the mechanism of synaptic potentiation in that a small postsynaptic effect was observed in glutamatergic synapses that was not seen at GABAergic synapses. Adenosine varied in the level of endogenous tonic inhibition that was present. Thus the major conclusion of this study is that different synapses in the NAcc are differentially modulated by these signaling pathways.

The selection of cells used in these recordings is an important consideration when interpreting these results. Cells were chosen from two subregions that were well within the boundaries of what is commonly considered the core and the shell of the NAcc (Paxinos and Watson 1998). Although MSNs are somewhat homogeneous in their physiological characteristics, neurochemical markers show complex staining patterns that suggest the NAcc has a complex and heterogeneous organization (Meredith 1999; Pennartz et al. 1994). This may reflect differences in interneurons, receptor distributions, synaptic inputs, and peptide cotransmitters. It is clear that we only sampled a small subset of the cells within two discrete regions of the NAcc, and other subdivisions were not considered. Hence there may be additional regional differences that are not reflected by our limited sample, and there may be heterogeneity within the regions from which we selected our neurons. Further selection may be caused by the placement of the stimulating electrodes. The local stimulation protocol used cannot identify which inputs are activated and only activates a subset of all available inputs. Hence the generalizability of these experiments is restricted by the small sample of cells and synapses studied. These subsets were made as consistent as possible through the placement of the electrodes (Fig. 1). Finally, the recording techniques used were optimal for detecting presynaptic changes in glutamate and GABA release, as assessed by measuring isolated AMPA and GABA_A receptor currents. Responses specific to NMDA receptors were not measured due to the blockade of these receptors. By limiting the study to a consistent set of responses, those responses could be directly compared across four different sets of synapses. However, these limitations must be taken into account when interpreting the data.

Opioid effects

The inhibitory effects of opioids were examined in the most detail on EPSCs in the shell. ME is a mixed µ- and -receptor agonist with no significant effect at -opioid receptors (Goldstein and Naidu 1989). Hence selective µ- and -receptor agonists were tested. The µ-receptor selective agonist DAMGO produced the same level of inhibition as ME, an effect that was reversible by the µ-selective antagonist CTAP. Furthermore, application of the -receptor selective agonist DPDPE failed to produce a statistically significant inhibition of these responses. It thus appears that µ-opioid receptors are the primary mediator of the effects observed, although a small inconsistent inhibition by -receptors cannot be ruled out. Interestingly, Yuan et al. (1992) found both µ and  inhibition of excitatory postsynaptic potentials recorded from MSNs in the NAcc core. In these experiments, the inhibition by µ-opioid receptors was largest, but the inhibition by -receptors was significant at higher stimulation intensities. It is possible that this is due to differences in recording parameters such as MSN selection and stimulating electrode placement that may alter the subpopulation of glutamate terminals that are being assayed. The EPSPs recorded by Yuan et al. (1992) contained mixed NMDA and/or AMPA receptor responses. Because we recorded isolated AMPA receptor EPSCs, it is also possible that -receptors preferentially inhibit NMDA receptor responses. In any case, our results confirm their findings that µ-opioid receptors produce the largest and most consistent inhibition of glutamate inputs to NAcc MSNs.

Shell IPSCs were inhibited by ME to a greater extent than were the other synapses. This suggests there may be differences in the expression and/or coupling of opioid receptors between different nerve terminals. The large effect of ME on shell IPSCs, coupled with the relatively small effect on shell EPSCs, suggests that the net effect of ME may be more excitatory (disinhibitory) in the shell than in the core. A recent study by Hoffman and Lupica (2001) found that DAMGO inhibited EPSCs in shell MSNs by a presynaptic mechanism but had no significant effect on IPSCs. However, the mixed µ- and -receptor agonist D-Ala²-Met⁵-enkephalinamide (DALA) produced a significant inhibition of IPSCs in ~50% of cells, leading the authors to conclude that GABAergic synapses were not inhibited by µ-receptors but that a subset was inhibited by -receptors. The present study found all of the GABAergic synapses tested were inhibited by the mixed µ- and/or -agonist ME. The data of Hoffman and Lupica (2001) suggest this may be a primarily -receptor-mediated effect. However, differences in the recording and stimulating locations make a direct comparison between these two studies difficult, and this may account for the more consistent inhibition of IPSCs observed in the present study.

Forskolin effects

Forskolin potentiated synaptic currents to a similar degree at all four synapses tested. Analysis of the paired-pulse ratio and miniature synaptic currents suggest that most of this potentiation was mediated by presynaptic effects. However, the smaller change in mEPSC amplitude suggests that there was a small postsynaptic effect of forskolin on shell MSN AMPA receptors. This is consistent with previous studies in which both presynaptic (Carroll et al. 1998; Chavez-Noriega and Stevens 1994; Chen and Regehr 1997) and postsynaptic (Greengard et al. 1991) enhancements in EPSCs were described. Interestingly, this potential postsynaptic effect was not observed in shell mIPSCs, suggesting there was no effect of forskolin on MSN GABA_A receptors. This is somewhat surprising, given the evidence that cAMP and/or PKA can modulate GABA_A receptor currents postsynaptically (McDonald et al. 1998; Poisbeau et al. 1999). However, there is clear evidence that IPSCs can be modulated by PKA through a presynaptic mechanism (Chieng and Williams 1998; Kondo and Marty 1997), and this appears to be the only site of action involved at this particular set of GABAergic synapses under these recording conditions. It is important to note, however, that it is possible some postsynaptic effects were missed due to washout of cAMP caused by the whole cell recording protocol. It cannot be conclusively stated that there are no cAMP- or PKA-mediated postsynaptic effects on GABA_A receptors under more native conditions, although there is a clear difference between AMPA and GABA_A receptors under these recording conditions.

Adenosine

The level of inhibition mediated by endogenous adenosine was determined by measuring the effects of the A₁ adenosine receptor antagonist DPCPX, which blocks the action of endogenous adenosine on A₁ receptors (Brundege and Dunwiddie 1996, 1998). DPCPX produced a significant increase in core EPSCs and in core and shell IPSCs, indicating that there was a significant tonic level of inhibition at these synapses, similar to that observed in other regions of the brain (Brundege and Dunwiddie 1996; Chieng and Williams 1998; Dunwiddie and Diao 1994; Manzoni et al. 1998). In contrast, DPCPX had no significant effect on shell EPSCs, indicating an extremely low level of tonic adenosine. This did not appear to be due to an insensitivity of these synapses to A₁ adenosine receptors, because both exogenously applied adenosine and the A₁ agonist CPA potently and effectively inhibited shell EPSCs. The most probable explanation is that the concentration of endogenous adenosine varies between the synapses, although differences in the receptors or signal transduction cannot be ruled out without a more detailed pharmacological comparison between the synapses. The possibility that adenosine concentrations vary between synapses that are interspersed and in close proximity (GABA and glutamate synapses in the shell) is intriguing. Adenosine is considered to be a "local hormone" or paracrine signaling agent rather than a neurotransmitter, and adenosine levels are thought to vary gradually over a relatively large spatial area (Arch and Newsholme 1978; Brundege and Dunwiddie 1997; Porkka-Heiskanen et al. 1997). If the concentration of adenosine is significantly different between different types of synapses on the same cell, it may suggest a more locally regulated mechanism than previously thought.

Conclusion

This study examined the effects of several major signaling pathways on excitatory synapses in the NAcc shell and compared the effects of these pathways between excitatory and inhibitory synapses in the core and shell regions of the NAcc. Each pathway studied showed some variation between the different synapses, either in maximal effect, mechanism of action, or concentration of neurohormone. These variations may reflect differences in the presynaptic terminals that synapse onto MSNs in the NAcc and may influence the relative sensitivity of these synapses to the effects of drugs acting on these systems.